Thu Sep 7 09:03:12 CEST 2006 [EMAIL PROTECTED]
* Use array-total-size-limit as the max size for strings
Thu Sep 7 09:03:39 CEST 2006 [EMAIL PROTECTED]
* Allow ub8 arrays to be autoconverted to foreign strings
New patches:
[Use array-total-size-limit as the max size for strings
[EMAIL PROTECTED] {
hunk ./src/strings.lisp 45
-(defun foreign-string-to-lisp (ptr &optional (size most-positive-fixnum)
+(defun foreign-string-to-lisp (ptr &optional (size array-total-size-limit)
}
[Allow ub8 arrays to be autoconverted to foreign strings
[EMAIL PROTECTED] {
hunk ./src/strings.lisp 40
- (loop with i = 0 for char across string
- while (< i size)
- do (%mem-set (char-code char) ptr :unsigned-char (post-incf i))
- finally (%mem-set 0 ptr :unsigned-char i)))
+ (etypecase string
+ (string
+ (loop with i = 0 for char across string
+ while (< i size)
+ do (%mem-set (char-code char) ptr :unsigned-char (post-incf i))
+ finally (%mem-set 0 ptr :unsigned-char i)))
+ ((array (unsigned-byte 8))
+ (loop with i = 0 for elt across string
+ while (< i size)
+ do (%mem-set elt ptr :unsigned-char (post-incf i))
+ finally (%mem-set 0 ptr :unsigned-char i)))))
hunk ./src/strings.lisp 68
- (check-type string string)
+ (check-type string (or string (array (unsigned-byte 8))))
hunk ./src/strings.lisp 105
- (if (pointerp obj)
- (values obj nil)
- (error "~A is not a Lisp string or pointer." obj)))
+ (cond
+ ((pointerp obj)
+ (values obj nil))
+ ((typep obj '(array (unsigned-byte 8)))
+ (values (foreign-string-alloc obj) t))
+ (t (error "~A is not a Lisp string, (array (unsigned-byte 8) or pointer." obj))))
hunk ./src/strings.lisp 140
- (if (pointerp obj)
- (values obj nil)
- (error "~A is not a Lisp string or pointer." obj)))
+ (cond
+ ((pointerp obj)
+ (values obj nil))
+ ((typep obj '(array (unsigned-byte 8)))
+ (values (foreign-string-alloc obj) t))
+ (t (error "~A is not a Lisp string, (array (unsigned-byte 8) or pointer." obj))))
hunk ./tests/misc-types.lisp 38
+
+(deftest misc-types.string+ptr.ub8
+ (destructuring-bind (string pointer)
+ (strdup (make-array 3 :element-type '(unsigned-byte 8) :initial-contents (map 'list #'char-code "foo")))
+ (foreign-free pointer)
+ string)
+ "foo")
}
Context:
[Use the RTEST package nickname instead of RT because of Lispworks.
Luis Oliveira <[EMAIL PROTECTED]>**20060906114027]
[Backwards compatibility for cffi-clisp's foreign-funcall
Luis Oliveira <[EMAIL PROTECTED]>**20060906101336]
[Support for recent CVS versions of CLISP
Luis Oliveira <[EMAIL PROTECTED]>**20060905103418
- As of 2006-09-03, CLISP's FFI::FOREIGN-LIBRARY-FUNCTION takes
an additional argument. Accomodated this change.
Patch courtesy of Kevin Rosenberg.
]
[[u]llong alignment issues on darwin/ppc
Luis Oliveira <[EMAIL PROTECTED]>**20060905101150
- CLISP: special case for :[unsigned-]long-long on darwin/ppc.
- SBCL: missing special case for :unsigned-long-long.
- New regression test: STRUCT.ALIGNMENT.8.
]
[New TODO items about LOAD-FOREIGN-LIBRARY.
Luis Oliveira <[EMAIL PROTECTED]>**20060831153130]
[New macro: INCF-POINTER (with documentation)
Luis Oliveira <[EMAIL PROTECTED]>**20060831153039]
[Misc manual cleanups like s/=>/@result{}/, etc...
Luis Oliveira <[EMAIL PROTECTED]>**20060831152837]
[Update cffi-uffi-compat to reflect recent UFFI changes.
Luis Oliveira <[EMAIL PROTECTED]>**20060831152423
- Better handling of shared library extensions on Windows.
Patch courtesy of Kevin Rosenberg.
]
[Add long-long support to CLISP
Luis Oliveira <[EMAIL PROTECTED]>**20060627011858
Patch courtesy of Frédéric Jolliton.
]
[ECL fixes
Luis Oliveira <[EMAIL PROTECTED]>**20060627010520
- cffi-ecl fixes:
- push cffi-features:unix on darwin too.
- use si:load-foreign-module instead of ffi:load-foreign-library
on DFFI platforms.
- use convert-external-name in foreign-symbol-pointer.
- wrap defcvar's define-foreign-symbol around eval-when.
- simplify default-library-suffix
- cffi-tests:
- load libtest.o on ECL platforms without DFFI.
- conditionalize the tests with 127 arguments based on the
value of lambda-parameters-limit.
]
[review Foreign Types; add Glossary
Stephen Compall <[EMAIL PROTECTED]>**20060613205550]
[uffi-compat: in load-foreign-library, don't probe for file if no directory.
Nathan Bird <[EMAIL PROTECTED]>**20060517201239
If it is just a filename without a directory, continue passing it down
to the underlying functions, as they probably know how to find a library
in default location.s
The test now matches the uffi behaviour too.
]
[Run tests both compiled and uncompiled.
Luis Oliveira <[EMAIL PROTECTED]>**20060607183000
- Make the test-suite run both with and without rt::*compile-tests*
bound to T.
]
[Huh. Found a very old and incomplete sentence in the manual.
Luis Oliveira <[EMAIL PROTECTED]>**20060607022358]
[Minor comestic change in foreign-vars.lisp
Luis Oliveira <[EMAIL PROTECTED]>**20060607022335]
[Lispworks bugfix: %mem-ref and %mem-set compiler macros
Luis Oliveira <[EMAIL PROTECTED]>**20060607022325
- The %mem-ref and %mem-set in cffi-lispworks.lisp were using
bogus indexes. FLI's documentation suggests foreign-typed-aref
expects array indexes but it seems to want offsets in bytes
instead.
- Regression tests: mem-ref.rt.1 and mem-ref.rt.2.
]
[bugfix: accept symbols in defcvar
Luis Oliveira <[EMAIL PROTECTED]>**20060527020411
- Fix lisp-var-name to accept symbols.
- Regression test: foreign-globals.symbol-name
]
[Minor simplification in tests/bindings.lisp
Luis Oliveira <[EMAIL PROTECTED]>**20060526113626]
[Fix some ECL bugs (maybe)
Luis Oliveira <[EMAIL PROTECTED]>**20060526113437
- Fixed some bitrot in cffi-ecl.lisp (maybe). Still can't test
properly because of an ECL bug related to make-load-form.
]
[Fix uffi-compat bugs
Luis Oliveira <[EMAIL PROTECTED]>**20060520181344
- :pointer is no longer a built-in type so we have to parse it,
not find-type it.
- def-array-pointer actually defines an array type with 1 element.
(this emulates UFFI's behaviour)
Bug report and initial patches courtesy of Lou Vanek.
]
[Fix problem with declarations in DEFCALLBACKs
Luis Oliveira <[EMAIL PROTECTED]>**20060514234218
- Place declarations after the translations take place.
- We no longer poke at ignore declarations in defcallback.
]
[change first automatic defbitfield value to 1
Stephen Compall <[EMAIL PROTECTED]>**20060512153832
- make-foreign-bitfield now starts implicit bitfield values at 1. No
more special cases for zero.
- Document and update bf tests to match.
]
[defbitfield: explicit initial zero case
Stephen Compall <[EMAIL PROTECTED]>**20060511162202
- Provide for counting an initial zero as a pseudo-single-bit.
- Test bitfield.4.
]
[only single-bits affect defbitfield implicit values
Stephen Compall <[EMAIL PROTECTED]>**20060511154221
- Remove a case in which a non-single-bit like 3 could hijack the
implicit value computation for defbitfield.
- New test bitfield.3 for this behavior.
]
[Fix defbitfield bug and new test
Luis Oliveira <[EMAIL PROTECTED]>**20060511145035
- Bug fix: when the first value was provided a (< foo nil) comparision
would occur. Regresion test: bitfield.1.
- New test bitfield.2.
]
[implicit defbitfield symbol values
Stephen Compall <[EMAIL PROTECTED]>**20060511072106
- Foreign Type Translators: defctype does not create Lisp types; you
have to use eql specializers.
- enum.lisp: Use reduce in %foreign-bitfield-value. Code a default
rule for bitfield symbol values.
]
[Make :POINTER a parameterized foreign type.
James Bielman <[EMAIL PROTECTED]>**20060507003511
- :POINTER without arguments is a void pointer.
- (:POINTER :INT) is a pointer to an :INT.
- These nest properly: (:POINTER (:POINTER :INT)).
- Pointers are not type checked yet---an optional pointer type checker
will be added someday.
]
[Bugfix: Quote type arguments to %DEFCFUN-VARARGS.
James Bielman <[EMAIL PROTECTED]>**20060507002829]
[Recognize OpenMCL/X86-64 and set CFFI features accordingly.
James Bielman <[EMAIL PROTECTED]>**20060504173940]
[Conditionalize against non-CPU specific word-size features in OpenMCL.
James Bielman <[EMAIL PROTECTED]>**20060504021623]
[Bugfix: Specialize UNPARSE for UFFI-CHAR in CFFI-UFFI-COMPAT.
James Bielman <[EMAIL PROTECTED]>**20060504015421
- Fixes an error when loading FASL files that dumped UFFI-CHAR instead
of (UFFI-CHAR :CHAR) using MAKE-LOAD-FORM.
- Reported by Ricardo Boccato.
]
[More copyright header year updates.
James Bielman <[EMAIL PROTECTED]>**20060503064225]
[Update copyright year in file headers.
James Bielman <[EMAIL PROTECTED]>**20060503063944]
[Conditionally set variables for implementations in Makefile.
James Bielman <[EMAIL PROTECTED]>**20060503063241
- Allow overriding from the environment as suggest in the comment, which
didn't actually work. (eg.: OPENMCL=openmcl64 make test)
]
[Add OpenMCL/X86-64 fasl files to 'make clean'.
James Bielman <[EMAIL PROTECTED]>**20060503061637]
[Make corman's finalizers thread-safe (hopefully)
Luis Oliveira <[EMAIL PROTECTED]>**20060424173620]
[Add support for finalizers
Luis Oliveira <[EMAIL PROTECTED]>**20060424025357
- New functions: finalize and cancel-finalization.
- New cffi-"feature": no-finalizers. Only ECL pushes this.
- Document new functions.
]
[Include stdint.h in libtest.c
Luis Oliveira <[EMAIL PROTECTED]>**20060424025320]
[Oops, forgot to bump the version number.
Luis Oliveira <[EMAIL PROTECTED]>**20060424025119]
[TAG 0.9.1
Luis Oliveira <[EMAIL PROTECTED]>**20060418011351]
Patch bundle hash:
846f09e9fbbb9728d9272336e73d2cc390325ce5
_______________________________________________
cffi-devel mailing list
cffi-devel@common-lisp.net
http://common-lisp.net/cgi-bin/mailman/listinfo/cffi-devel
